Skip to main content

Table 3 Housing and homelessness outcomes

From: Interventions to improve social circumstances of people with mental health conditions: a rapid evidence synthesis

Mental health diagnoses

Author

Intervention vs Control

Badged as 'Housing First'

Independent tenancy

Staff on site/supported housing

multi-disciplinary team mental health support (e.g. ACT, ICM)

Housing support worker practical support (no multi-disciplinary team support)

Specified psychological therapy offered

Outcomes

 

Achieving/Sustaining Housing

SMI

Aubry 2016 (Canada, N = 950) [58]

Housing First + ACT vs TAU

Y

Y

N

Y

N

N

Housing First participants spent a significantly higher percentage of time in stable housing compare to the control group at 6 months (Hedges g = 1.41, 95% CI: 1.26, 1.55), 12 months (Hedges g = 1.14, 95% CI: 1.00, 1.27) and 24 months (Hedges g = 0.59, 95% CI: 0.46, 0.72) post baseline

 

Tinland 2020 (France, N = 703) [59]

Housing First + ACT vs treatment as usual

Y

Y

N

Y

N

N

The Housing First + ACT group had significantly more mean days stably housed at 6 months from baseline (Hedges g = 2.22, 95% CI: 2.04, 2.41), 12 months from baseline (Hedges g = 1.81, 95% CI: 1.64, 1.99) and 24 months from baseline (Hedges g = 1.38, 95% CI: 1.21, 1.54)

 

Stergiopoulos 2015 (Canada, N = 1198) [60]

Housing First plus integrated case management vs treatment as usual

Y

Y

N

Y

N

N

The Housing First plus integrated case management group spent a significantly higher mean number of days stably housed than the treatment as usual group at 24 months (Hedges g = 0.85, 95% CI: 0.73, 0.97)

 

Tsemberis 2004 (USA, N = 206) [61]

Pathways Housing First vs continuum of care

Y

Y

N

Y

N

N

a repeated measures ANOVA showed a time x group status effect such that participants in the experimental condition had significantly faster decreases in homelessness (F(4, 137) = 10.1, P < .001) and increases in stable housing (F(4, 137) = 27.7, P < .001) relative to control participants

 

Burnam 1996 (USA, N = 276) [62]

a) Social model residential treatment programme

b) Social model non-residential treatment programme

vs no intervention

a) N

b) N

a) N

b) N

a) Y

b) N

a) Y

b) Y

a) N

b) N

a) N

b) N

There were no significant differences between residential and non-residential housing interventions, except that non-residential housing participants were more likely to have increased the amount of time they spent in independent housing at 3 months following baseline. this is expected because residential housing participants were by definition not stably housed

 

Ellison 2020 (USA, N = 166) [63]

Manualized treatment model for co-occurring mental illness and substance use disorders (MISSION-Vet) vs TAU

N

N

N

N

N

Y

Treated veterans did not spend more days in housing compared with control veterans during any part of the study at 95% level of confidence

 

Fletcher 2008 (USA, N = 161) [64]

Integrated ACT vs standard care

N

N

N

Y

N

Y

There was no significant difference between the integrated Assertive Community Treatment and the control group in the mean number of days stably housed (15 months: Hedges g = 0.25, 95% CI: -0.10, 0.61, 30 months: Hedges g = 0.31, 95% CI: -0.04, 0.66)

However, the following variables mediated the IACT vs. SC effect on stable housing: programme contacts, substance abuse contacts, help with activities of daily living, transportation assistance, and, help with medication

 

Goldfinger 1999 (USA, N = 118) [65]

Group housing vs independent housing

N

N

Y

Y

N

N

The intervention group were not significantly more likely to be in stable housing than independent housing control group at the end of 18-month follow-up period (OR = 1.09, 95% CI: 0.45, 2.63)

 

Herman 2011 (USA, N = 150) [66]

Critical time intervention + usual care vs usual care

N

N

N

Y

N

N

At 18 months, the OR of experiencing homelessness in the intervention group compared to the control group during the final three follow up intervals was 0.22 (95% CI:0.06, 0.88) when controlling for baseline homelessness, indicating that the intervention group were less likely to experience homelessness

 

Hurlburt 1996 (USA, N = 361) [67]

a) Sect. 8 rent subsidy certificate vs no Sect. 8 certificate

b) Comprehensive housing services vs traditional housing services

a) N

b) N

a) N

b) N

a) N

b) N

a) N

b) Y

a) Y

b) N

a) N

b) N

a) a very strong relationship appeared between Sect. 8 housing condition and housing outcomes. Most of this effect was on type of community housing obtained- of those who achieved stable housing, those on Sect. 8 were 4.87 times likely to achieve stability in independent housing than those in the non-Sect. 8 condition. However, when categories of independent housing and other consistent housing were combined, Sect. 8 clients were only 1.21 times more likely to achieve some type of consistent housing than non-Sect. 8 clients

Section 8 certificates were strongly associated with obtaining independent housing, regardless of substance abuse diagnosis

b) There was no difference in housing stability when traditional services were compared to comprehensive services

 

Korr & Joseph 1995 (USA, N = 95) [68]

Case management vs routine care

N

N

Y

N

Y

N

At 6 months post baseline, over twice as many experimental clients as control clients were housed (OR = 6.4, 95% CI: 2.61, 15.68)

 

Lehman 1997 (USA, N = 152) [69]

ACT vs usual community services

N

N

N

Y

N

N

The ACT group spent significantly more days in the 12-month follow-up housed (210.2 vs 160.1, Hedges g = 0.46, 95% CI: 0.14, 0.78)

 

Lipton 1988 (USA, N = 49) [70]

Residential treatment vs standard care

N

N

N

Y

N

N

More participants in the intervention group were in permanent housing at 12 months than the control group (OR = 5.19, 95% CI: 2.84, 9.48)

 

McHugo 2004 (USA, N = 121) [71]

Integrated housing vs Parallel housing

N

N

Y

Y

N

N

The integrated housing condition did not spend a significantly higher mean number of days in stable housing at 6 months (Hedges g = 0.20, 95% CI: -0.16, 0.55, P = 0.287), however, they spent significant more days in stable housing at both 12 months (Hedges g = 0.50, 95% CI: 0.14, 0.86) and 18 months (Hedges g = 0.50, 95% CI: 0.134, 0.86)

 

Morse 1992 (USA, N = 116) [72]

Continuous treatment team vs outpatient mental health services

N

N

N

Y

N

N

A significant treatment-by-time interaction was found for days homeless (F = 4.23, df = 2,97, P = .017). Post hoc analyses indicated that the clients in the continuous treatment team programme were less likely to be homeless at 12 months than those in the outpatient programme, however, endpoint analyses at 12 months did not show a significant group effect (Hedges g = 0.29, 95% CI: -0.17, 0.75)

 

Morse 2006 (USA, N = NR) [73]

Integrated ACT vs standard care

N

N

N

Y

N

Y

The main effect of treatment on stable housing was statistically significant at 24 months follow up, F(2, 145) = 3.76, p = .03. Post-hoc analyses indicated that clients in IACT condition had significantly more days in stable housing than control clients. There was no significant treatment by time interaction, F(6, 440) = 1.93, p = .07

 

Shern 2000 (USA, N = 168) [74]

Community outreach (Choices) vs Treatment as usual

N

N

Y

N

Y

N

Participants in the community outreach programme reported a 23% increase in the proportion of time spent in shelters (t = -5.73, P < .001) compared to the control group at 24 months

 

Susser 1997 (USA, N = 96) [75]

Critical time intervention vs usual services

N

N

N

N

Y

N

During the last month of the 18- month follow-up, 4 (8%) of the men in the CTI group and 11 (23%) of the men in the USO group were homeless, a significant difference (OR of not being homeless = 3.46, 95% CI: 1.01, 11.80, P = 0.047)

Mixed mental health conditions

Morse 1997 (USA, 165) [76]

a) Broker case management vs ACT

b) ACT with community workers

vs ACT

a) N

b) N

a) N

b) N

a) N

b) N

a) Y

b) Y

a) N

b) N

a) N

b) N

A significant treatment group effect was found on days in stable housing (F = 3.54, df = 2,129, P < 0.032), such that Assertive Community Treatment only participants averaged more days in stable housing at 18 months than clients in both broker case management and Assertive Community Treatment with community workers

 

Housing Quality

SMI

Aubry 2016 (Canada, N = 950) [58]

Housing First + ACT vs TAU

Y

Y

N

Y

N

N

Compared with treatment-as-usual participants, Housing First participants rated their housing as being of significantly better quality at 6 months (Hedges g = 0.60, 95% CI: 0.47, 0.73), 12 months (Hedges g = 0.57, 95% CI: 0.44, 0.70) and 24 months from baseline (Hedges g = 0.22, 95% CI: 0.09, 0.34)

Lehman 1997 (USA, N = 152) [69]

ACT vs usual community services

N

N

N

Y

N

N

The ACT programme subjects were significantly more satisfied with their housing at the 6-month follow-up (Hedges g = 0.37, 95% CI: 0.05, 0.68), but not at 12-month follow-up (Hedges g = 0.09, 95% CI: -0.26, 0.44)

Lipton 1988 (USA, N = 49) [70]

Residential treatment vs standard care

N

N

N

Y

N

N

The experimental group, with a mean score of 1.63 at 12 months, indicated that on average they were satisfied with and committed to their housing arrangements. At 12 months the controls’ mean rating of their living arrangements was 2.87, indicating that on average they perceived inadequacies and desired an alternative (Hedges g = 1.25, 95% CI: 0.50, 2.00)

  1. N Number of participants, SMI Severe mental illness, TAU Treatment as usual, CBT Cognitive behavioural training, ACT Assertive Community Treatment, OR Odds ratio, CI Confidence interval, Y Yes, N No